On Sun, Jan 6, 2013 at 9:19 PM, Salomon Turgman Cohen <[email protected]>wrote:

> Hello all,
>    After some experimentation it seems to me the values of facesphi do
> update after every time step and that the approach outlined in the second
> email is a valid one. If I attempt to use just phi as the variable used to
> set D I get an error due to the difference in shapes between D and phi. It
> seems that my second email outlines the correct method to use a coefficient
> that depends in a piecewise fashion on the variable being solved (were the
> coeff is defined at the cell faces). Please correct me if I am wrong.
>

Use the "cases" function I provided for piecewise operations. If you want D
to be a FaceVariable then you can either submit phi.faceValue to "cases" or
pass D.faceValue as the coefficient.



>    Would it be worthwhile to expand the mesh1D diffusion example to
> include a piecewise version of the final part of that example in which the
> diffusion coefficient depends linearly on concentration? I am willing to
> write a new version of the example and post it here if there is interest to
> make the change.
>

Sounds like a good idea. Thanks. We probably need a better solution to
piecewise dependency than the solution that I provided due to the inherent
inefficiencies with the "<" and "*" method. Not sure what that would be
though.

Cheers

-- 
Daniel Wheeler
_______________________________________________
fipy mailing list
[email protected]
http://www.ctcms.nist.gov/fipy
  [ NIST internal ONLY: https://email.nist.gov/mailman/listinfo/fipy ]

Reply via email to